metadata
dataset_info:
features:
- name: filename
dtype: string
- name: title
dtype: string
- name: abstract
dtype: string
- name: pred_lin
sequence: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 5002011
num_examples: 525
download_size: 1118062
dataset_size: 5002011
configs:
- config_name: default
data_files:
- split: train
path: data/train-*